As physics-based global magnetosphere-ionosphere-thermosphere models enter operations for space weather forecasting, it is imperative that the sources of uncertainty in model outputs are thoroughly investigated. Global magnetosphere models rely on a set of input solar wind conditions to drive the model; such inputs are often propagated from upstream in the solar wind where the IMF and plasma parameters can be constantly monitored. Generally, this data is provided by the OMNI dataset, a set of observations of the solar wind typically collected near the L1 Lagrange point and propagated to Earth’s bow shock nose. The actual solar wind conditions outside the bow shock may differ from the OMNI data, leading to erroneous model inputs. Erroneous model inputs will lead to erroneous model outputs, and contribute to poor predictions of space weather phenomena. We present a study on the effects of variability in OMNI solar wind propagation accuracy on ionosphere-thermosphere model outputs at the individual grid cell level. Models are driven with OMNI and ARTEMIS solar wind time series for three events at varying levels of correlation between both inputs (Pearson Correlation Coefficient = 0.95, 0.82, 0.58). Thermospheric neutral densities at three heights (200 km, 400 km, 600 km) are compared for each simulation at grid cells over Kennedy Space Center, Florida, USA and College Magnetometer Station, Alaska, USA.